Abstract: Step line is one of the main objects of open-pit mine survey. The traditional point measurement method of total station and GPS is time-consuming,labor-intensive and inefficient,while the modern measurement method based on UAV point cloud provides the possibility for efficient extraction of step lines. Based on the dense matching point cloud data constructed from UAV images,a step line extraction method with adaptive elevation threshold was proposed. Firstly,the terrain condition was judged by the covariance matrix of the neighborhood point cloud,and different elevation difference thresholds were adaptively selected according to different terrain to extract the step line feature points,which greatly reduced the number of false extraction. Then,based on the number of points in each point cloud cluster and elevation characteristics after point cloud clustering segmentation,the misextracted point clouds were removed,and the cubic B-spline curve was used to fit the feature points. Finally,the step line extraction of open-pit mine was completed. The test results showed that the method can efficiently extract the feature points of the step line in open-pit mine,and effectively distinguish the step line and the road edge line feature points. At the same time,the curve fitted according to the feature points has a good effect and truly reflects the actual position of the step line. The above analysis provides a new solution for the efficient measurement of step lines in open-pit mines.